1. Field of the Invention
The present invention relates generally to a system and method for iris recognition using stereoscopic face recognition, which can recognize irises using a stereoscopic face image in order to recognize an authenticatee.
2. Related Art
As well known to those skilled in the art, in biometrics technologies parts of a human body are considered as recognition objects to recognize a person, iris, face, fingerprint, vein and the like recognition technologies are independently utilized.
However, certain biometrics technologies cannot be applied to certain persons. For example, the iris recognition technology cannot be applied to persons who are unable to open their eyes wide by nature or due to acquired diseases. The fingerprint recognition technology cannot be applied to persons who have worn fingerprints. Further, the face recognition technology cannot be applied to persons whose faces have been changed by plastic surgery, and so on. Accordingly, combined biometrics technologies, instead of the various independent biometrics technologies, have recently been developed to overcome these defects.
The overall recognition accuracy of the combined biometrics technologies is generally higher than the independent biometrics technologies applied individually.
For example, it is assumed that recognition accuracies of biometrics technologies A and B, different from each other, are 90% and 80%, respectively. If the biometrics technologies A and B are applied to one hundred persons, the ninety persons are recognized through the biometrics technology A in the case where the biometrics technology A is applied to the hundred persons, and eight persons are additionally recognized through the biometrics technology B in the case where the biometrics technology B is applied to the rest ten persons not recognized by the biometrics technology A. Accordingly, the overall recognition accuracy is 98%.
When the overall recognition accuracy is expressed as a formula, the overall recognition accuracy in the case where independent biometrics technologies having independent recognition accuracies with a % and b % are applied to persons one by one, is represented as the following Formula:a+(100−a)*b/100(%).  [Formula]
The overall recognition accuracy is higher than the cases where the biometrics technologies A and B are independently applied.
Meanwhile, the conventional face recognition technology is utilized by taking a face image using a single camera, such that the face image cannot be analyzed in three dimensions. Accordingly, the face recognition technology has limitations, wherein the location of pupils can be detected but the distance information of the pupils from an iris camera cannot be obtained, both of which are required to capture the focused iris images.
Further, the conventional iris recognition technology has an added disadvantage in terms of the high cost auto-focusing camera of the iris recognition system. The conventional used, manually focused iris recognition camera system is inconvenient because it requires adjustment of the focal length and optical axis after aligning the pupils with the lens of the camera.
Moreover, the systems for iris recognition and face recognition are not operated as a module but are independently operated and so the systems are expensive and it is difficult to obtain a synergistic effect resulting from applying the location information of the pupils and the distance information of the pupils from a camera obtained in the process of the face recognition to iris recognition.
In a further improved system using a face recognition camera, one of the iris recognition cameras recognizes the locations of the pupils in the process of the face recognition, and then carries out the iris recognition. However, since a single face recognition camera is used, distances of the pupils from the iris recognition camera cannot be obtained, and a user must position his pupils at the optical axis of the camera so that an image is photographed in the vicinity of the focal point of the camera.